Awesome! It looks like we're onto something
@Thistleknot:
Is that a working multimerge? Wow.
Now there's the just the comparatively simple tasks of implementing the flatten and merge functions in Python, rejecting failed merges, and then opening it up! (Seriously, if the logic is confirmed to work that's a massive step forward)
@Button:
I think that given the novelty of easy patch based loading we might as well err on the side of many tiny patches for once - just to see what that's like! Medium term, I can imagine that some things might amalgamate a little; to take the example of Accelerated DF it might want to divide materials into 'Generic creatures mod' and 'Generic stone and wood mod', but splitting blood from leather from shell etc would be too much.
What are y'all's thoughts on applying graphics packs to the resulting raw monstrosity? (And I use the term affectionately). Are mods going to have to be included in graphics'd versions in order to be used with a tileset; will graphics packs be treated like just another mod, merged in at the end; or will there be special handling for graphics packs?
Ideally graphics will be treated like any other mod, meaning mods should be applied to vanilla, and the Mod Starter Pack would merge the mod with the graphics. It remains to be seen how practical that is.
Due to practical concerns, I'd lean the other way - treat graphics as a special case and assume that we're using ASCII with [graphics:no]; note that tilesets are still compatible, just not graphics.
Some graphics packs are based on standard raws and others aren't. This is historically to free up the tiles for accented letters for other things, but with the Text will be Text plugin that could be reversed - which would mean that the raw folder could be left entirely alone by graphics packs. They also have to mess around with the /data folder a lot, and if we can avoid needing to touch that we probably should - separation of concerns to avoid causing yet more conflicts. I'm aware that this limits the range of compatible mods somewhat, but I don't see encouraging a return to the baseline standard as a terrible position to take; and we can always extend later.